MMNavAgent: Multi-Magnification WSI Navigation Agent for Clinically Consistent Whole-Slide Analysis
🤖 AI Summary
This study addresses the limitation of existing AI methods in whole-slide image (WSI) diagnosis, which typically rely on a single or predefined magnification level and thus fail to emulate the multi-scale, dynamic interaction inherent in pathologists’ clinical workflow. To bridge this gap, the authors propose a clinically aligned multi-magnification navigation agent that innovatively integrates a Cross-Magnification Navigation Tool (CMT) and a Magnification Selection Tool (MST). By leveraging context-aware multi-scale representation fusion and memory-driven adaptive decision-making, the agent performs sequential diagnostic analysis akin to human pathologists. Evaluated on public datasets, the proposed method significantly outperforms non-agent baselines, achieving a 1.45% improvement in AUC and a 2.93% gain in balanced accuracy (BACC).
Application Category
📝 Abstract
Recent AI navigation approaches aim to improve Whole-Slide Image (WSI) diagnosis by modeling spatial exploration and selecting diagnostically relevant regions, yet most operate at a single fixed magnification or rely on predefined magnification traversal. In clinical practice, pathologists examine slides across multiple magnifications and selectively inspect only necessary scales, dynamically integrating global and cellular evidence in a sequential manner. This mismatch prevents existing methods from modeling cross-magnification interactions and adaptive magnification selection inherent to real diagnostic workflows. To these, we propose a clinically consistent Multi-Magnification WSI Navigation Agent (MMNavAgent) that explicitly models multi magnification interaction and adaptive magnification selection. Specifically, we introduce a Cross-Magnification navigation Tool (CMT) that aggregates contextual information from adjacent magnifications to enhance discriminative representations along the navigation path. We further introduce a Magnification Selection Tool (MST) that leverages memory-driven reasoning within the agent framework to enable interactive and adaptive magnification selection, mimicking the sequential decision process of pathologists. Extensive experiments on a public dataset demonstrate improved diagnostic performance, with 1.45% gain of AUC and 2.93% gain of BACC over a non-agent baseline. Code will be public upon acceptance.
Problem
Research questions and friction points this paper is trying to address.
Whole-Slide Image
multi-magnification
adaptive magnification selection
clinical diagnosis
cross-magnification interaction
Innovation
Methods, ideas, or system contributions that make the work stand out.
multi-magnification navigation
adaptive magnification selection
cross-magnification interaction
memory-driven reasoning
whole-slide image analysis
🔎 Similar Papers
No similar papers found.
